10 - ``gnt-instance batch-create`` has been changed to use the bulk create
11 opcode from Ganeti. This lead to incompatible changes in the format of
12 the JSON file. It's now not a custom dict anymore but a dict
13 compatible with the ``OpInstanceCreate`` opcode.
14 - Parent directories for file storage need now to be listed in
15 ``$sysconfdir/ganeti/file-storage-paths``. ``cfgupgrade`` will write
16 the file automatically based on old configuration values, but it can
17 not distribute it across all nodes and the file contents should be
18 verified. Use ``gnt-cluster copyfile
19 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
20 upgraded. The reason for requiring this list of paths now is that
21 before it would have been possible to inject new paths via RPC,
22 allowing files to be created in arbitrary locations. The RPC protocol
23 is protected using SSL/X.509 certificates, but as a design principle
24 Ganeti does not permit arbitrary paths to be passed.
25 - The parsing of the variants file for OSes (see
26 :manpage:`ganeti-os-interface(8)` has been slightly changed: now empty
27 lines and comment lines are ignored for better readability.
33 *(Released Fri, 12 Oct 2012)*
35 A small bugfix release.
37 Fix double use of PRIORITY_OPT in gnt-node migrate, that would make the
40 Commands that issue many jobs don't fail anymore just because some jobs
41 take so long that other jobs are archived.
43 Failures during gnt-instance reinstall are reflected by the exit status.
45 Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
48 When always_failover is set, --allow-failover is not required in migrate
51 bash_completion works even if extglob is disabled
53 Fix bug with locks that made failover for RDB-based instances fail.
55 Fix bug in non-mirrored instance allocation that would make Ganeti
56 choose a random node instead of one based on the allocator metric.
58 Support for newer versions of pylint and pep8.
60 Hail doesn't fail anymore when trying to add an instance of type
61 'file', 'sharedfile' or 'rbd'.
63 Add new Makefile target to rebuild the whole dist, so that all files are
70 *(Released Fri, 27 Jul 2012)*
73 .. attention:: The ``LUXI`` protocol has been made more consistent
74 regarding its handling of command arguments. This, however, leads to
75 incompatibility issues with previous versions. Please ensure that you
76 restart Ganeti daemons soon after the upgrade, otherwise most
77 ``LUXI`` calls (job submission, setting/resetting the drain flag,
78 pausing/resuming the watcher, cancelling and archiving jobs, querying
79 the cluster configuration) will fail.
88 The current ``admin_up`` field, which used to denote whether an instance
89 should be running or not, has been removed. Instead, ``admin_state`` is
90 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
92 The rational behind this is that an instance being “down” can have
95 - it could be down during a reboot
96 - it could be temporarily be down for a reinstall
97 - or it could be down because it is deprecated and kept just for its
100 The previous Boolean state was making it difficult to do capacity
101 calculations: should Ganeti reserve memory for a down instance? Now, the
102 tri-state field makes it clear:
104 - in ``up`` and ``down`` state, all resources are reserved for the
105 instance, and it can be at any time brought up if it is down
106 - in ``offline`` state, only disk space is reserved for it, but not
109 The field can have an extra use: since the transition between ``up`` and
110 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
111 transition between ``offline`` and ``down`` is done via ``gnt-instance
112 modify``, it is possible to given different rights to users. For
113 example, owners of an instance could be allowed to start/stop it, but
114 not transition it out of the offline state.
116 Instance policies and specs
117 +++++++++++++++++++++++++++
119 In previous Ganeti versions, an instance creation request was not
120 limited on the minimum size and on the maximum size just by the cluster
121 resources. As such, any policy could be implemented only in third-party
122 clients (RAPI clients, or shell wrappers over ``gnt-*``
123 tools). Furthermore, calculating cluster capacity via ``hspace`` again
124 required external input with regards to instance sizes.
126 In order to improve these workflows and to allow for example better
127 per-node group differentiation, we introduced instance specs, which
130 - minimum instance disk size, disk count, memory size, cpu count
131 - maximum values for the above metrics
132 - and “standard” values (used in ``hspace`` to calculate the standard
135 The minimum/maximum values can be also customised at node-group level,
136 for example allowing more powerful hardware to support bigger instance
139 Beside the instance specs, there are a few other settings belonging to
140 the instance policy framework. It is possible now to customise, per
141 cluster and node-group:
143 - the list of allowed disk templates
144 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
145 - the maximum ratio of instance to spindles (see below for more
146 information) for local storage
148 All these together should allow all tools that talk to Ganeti to know
149 what are the ranges of allowed values for instances and the
150 over-subscription that is allowed.
152 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
153 instance configuration, and the physical CPU configuration from the
154 node. For the spindle ratios however, we didn't track before these
155 values, so new parameters have been added:
157 - a new node parameter ``spindle_count``, defaults to 1, customisable at
158 node group or node level
159 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
161 Note that spindles in this context doesn't need to mean actual
162 mechanical hard-drives; it's just a relative number for both the node
163 I/O capacity and instance I/O consumption.
165 Instance migration behaviour
166 ++++++++++++++++++++++++++++
168 While live-migration is in general desirable over failover, it is
169 possible that for some workloads it is actually worse, due to the
170 variable time of the “suspend” phase during live migration.
172 To allow the tools to work consistently over such instances (without
173 having to hard-code instance names), a new backend parameter
174 ``always_failover`` has been added to control the migration/failover
175 behaviour. When set to True, all migration requests for an instance will
176 instead fall-back to failover.
178 Instance memory ballooning
179 ++++++++++++++++++++++++++
181 Initial support for memory ballooning has been added. The memory for an
182 instance is no longer fixed (backend parameter ``memory``), but instead
183 can vary between minimum and maximum values (backend parameters
184 ``minmem`` and ``maxmem``). Currently we only change an instance's
187 - live migrating or failing over and instance and the target node
188 doesn't have enough memory
189 - user requests changing the memory via ``gnt-instance modify
195 In order to control the use of specific CPUs by instance, support for
196 controlling CPU pinning has been added for the Xen, HVM and LXC
197 hypervisors. This is controlled by a new hypervisor parameter
198 ``cpu_mask``; details about possible values for this are in the
199 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
200 VCPU-to-CPU mapping) form will work well only when all nodes in your
201 cluster have the same amount of CPUs.
206 Another area in which Ganeti was not customisable were the parameters
207 used for storage configuration, e.g. how many stripes to use for LVM,
208 DRBD resync configuration, etc.
210 To improve this area, we've added disks parameters, which are
211 customisable at cluster and node group level, and which allow to
212 specify various parameters for disks (DRBD has the most parameters
213 currently), for example:
215 - DRBD resync algorithm and parameters (e.g. speed)
216 - the default VG for meta-data volumes for DRBD
217 - number of stripes for LVM (plain disk template)
220 These parameters can be modified via ``gnt-cluster modify -D …`` and
221 ``gnt-group modify -D …``, and are used at either instance creation (in
222 case of LVM stripes, for example) or at disk “activation” time
225 Rados block device support
226 ++++++++++++++++++++++++++
228 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
229 denoted by the ``rbd`` disk template type. This is considered
230 experimental, feedback is welcome. For details on configuring it, see
231 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
236 The existing master IP functionality works well only in simple setups (a
237 single network shared by all nodes); however, if nodes belong to
238 different networks, then the ``/32`` setup and lack of routing
239 information is not enough.
241 To allow the master IP to function well in more complex cases, the
242 system was reworked as follows:
244 - a master IP netmask setting has been added
245 - the master IP activation/turn-down code was moved from the node daemon
247 - whether to run the Ganeti-supplied master IP script or a user-supplied
248 on is a ``gnt-cluster init`` setting
250 Details about the location of the standard and custom setup scripts are
251 in the man page :manpage:`gnt-cluster(8)`; for information about the
252 setup script protocol, look at the Ganeti-supplied script.
257 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
260 It is now possible to use TLS-protected connections, and when renewing
261 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
262 it is now possible to specify spice or spice CA certificates. Also, it
263 is possible to configure a password for SPICE sessions via the
264 hypervisor parameter ``spice_password_file``.
266 There are also new parameters to control the compression and streaming
267 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
268 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
269 for the spice parameters.
271 Lastly, it is now possible to see the SPICE connection information via
272 ``gnt-instance console``.
277 A new tool (``tools/ovfconverter``) has been added that supports
278 conversion between Ganeti and the `Open Virtualization Format
279 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
282 This relies on the ``qemu-img`` tool to convert the disk formats, so the
283 actual compatibility with other virtualization solutions depends on it.
288 The configuration query daemon (``ganeti-confd``) is now optional, and
289 has been rewritten in Haskell; whether to use the daemon at all, use the
290 Python (default) or the Haskell version is selectable at configure time
291 via the ``--enable-confd`` parameter, which can take one of the
292 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
293 daemon will result in a smaller footprint; for larger systems, we
294 welcome feedback on the Haskell version which might become the default
297 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
298 daemon running. The Python version doesn't implement the new call.
301 User interface changes
302 ~~~~~~~~~~~~~~~~~~~~~~
304 We have replaced the ``--disks`` option of ``gnt-instance
305 replace-disks`` with a more flexible ``--disk`` option, which allows
306 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
307 disk size and mode can be changed upon recreation (via ``gnt-instance
308 recreate-disks``, which accepts the same ``--disk`` option).
310 As many people are used to a ``show`` command, we have added that as an
311 alias to ``info`` on all ``gnt-*`` commands.
313 The ``gnt-instance grow-disk`` command has a new mode in which it can
314 accept the target size of the disk, instead of the delta; this can be
315 more safe since two runs in absolute mode will be idempotent, and
316 sometimes it's also easier to specify the desired size directly.
318 Also the handling of instances with regard to offline secondaries has
319 been improved. Instance operations should not fail because one of it's
320 secondary nodes is offline, even though it's safe to proceed.
322 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
323 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
324 minors to instance name.
329 RAPI coverage has improved, with (for example) new resources for
330 recreate-disks, node power-cycle, etc.
335 There is partial support for ``xl`` in the Xen hypervisor; feedback is
338 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
339 whether to still support Python 2.4 or move to Python 2.6 as minimum
342 Support for Fedora has been slightly improved; the provided example
343 init.d script should work better on it and the INSTALL file should
344 document the needed dependencies.
349 The deprecated ``QueryLocks`` LUXI request has been removed. Use
350 ``Query(what=QR_LOCK, ...)`` instead.
352 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
353 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
354 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
355 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
356 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
358 RAPI client: ``CertificateError`` now derives from
359 ``GanetiApiError``. This should make it more easy to handle Ganeti
362 Deprecation warnings due to PyCrypto/paramiko import in
363 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
364 make sure to run an up-to-date paramiko version, if you use this tool.
366 The QA scripts now depend on Python 2.5 or above (the main code base
367 still works with Python 2.4).
369 The configuration file (``config.data``) is now written without
370 indentation for performance reasons; if you want to edit it, it can be
371 re-formatted via ``tools/fmtjson``.
373 A number of bugs has been fixed in the cluster merge tool.
375 ``x509`` certification verification (used in import-export) has been
376 changed to allow the same clock skew as permitted by the cluster
377 verification. This will remove some rare but hard to diagnose errors in
384 *(Released Thu, 19 Jul 2012)*
386 Very few changes from rc4 to the final release, only bugfixes:
388 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
389 instance, fix CDROM booting for KVM instances)
390 - fixed node group modification of node parameters
391 - fixed issue in LUClusterVerifyGroup with multi-group clusters
392 - fixed generation of bash completion to ensure a stable ordering
399 *(Released Fri, 13 Jul 2012)*
401 Third release candidate for 2.6. The following changes were done from
404 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
405 - Fixed an inconsistency in the LUXI protocol with the provided
406 arguments (NOT backwards compatible)
407 - Fixed a bug with node groups ipolicy where ``min`` was greater than
408 the cluster ``std`` value
409 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
410 easier instance debugging on nodes (requires ``hconfd`` to work)
416 *(Released Tue, 03 Jul 2012)*
418 Second release candidate for 2.6. The following changes were done from
421 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
423 - Fixed a RAPI regression on missing beparams/memory
424 - Fixed redistribution of files on offline nodes
425 - Added possibility to run activate-disks even though secondaries are
426 offline. With this change it relaxes also the strictness on some other
427 commands which use activate disks internally:
428 * ``gnt-instance start|reboot|rename|backup|export``
429 - Made it possible to remove safely an instance if its secondaries are
431 - Made it possible to reinstall even though secondaries are offline
437 *(Released Mon, 25 Jun 2012)*
439 First release candidate for 2.6. The following changes were done from
442 - Fixed bugs with disk parameters and ``rbd`` templates as well as
444 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
445 behaviour. It supports now the modify operation
446 - ``hcheck`` implemented to analyze cluster health and possibility of
447 improving health by rebalance
448 - ``hbal`` has been improved in dealing with split instances
454 *(Released Mon, 11 Jun 2012)*
456 Second beta release of 2.6. The following changes were done from beta2
459 - Fixed ``daemon-util`` with non-root user models
460 - Fixed creation of plain instances with ``--no-wait-for-sync``
461 - Fix wrong iv_names when running ``cfgupgrade``
462 - Export more information in RAPI group queries
463 - Fixed bug when changing instance network interfaces
464 - Extended burnin to do NIC changes
465 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
466 - Changed default for DRBD barriers
467 - Fixed DRBD error reporting for syncer rate
468 - Verify the options on disk parameters
470 And of course various fixes to documentation and improved unittests and
477 *(Released Wed, 23 May 2012)*
479 First beta release of 2.6. The following changes were done from beta1 to
482 - integrated patch for distributions without ``start-stop-daemon``
483 - adapted example init.d script to work on Fedora
484 - fixed log handling in Haskell daemons
485 - adapted checks in the watcher for pycurl linked against libnss
486 - add partial support for ``xl`` instead of ``xm`` for Xen
487 - fixed a type issue in cluster verification
488 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
491 Plus integrated fixes from the 2.5 branch:
493 - fixed ``kvm-ifup`` to use ``/bin/bash``
494 - fixed parallel build failures
495 - KVM live migration when using a custom keymap
501 *(Released Tue, 24 Jul 2012)*
503 A small bugfix release, with no new features:
505 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
506 different default shell (e.g. Debian, Ubuntu)
507 - fixed KVM startup and live migration with a custom keymap (fixes Issue
508 243 and Debian bug #650664)
509 - fixed compatibility with KVM versions that don't support multiple boot
510 devices (fixes Issue 230 and Debian bug #624256)
512 Additionally, a few fixes were done to the build system (fixed parallel
513 build failures) and to the unittests (fixed race condition in test for
514 FileID functions, and the default enable/disable mode for QA test is now
521 *(Released Fri, 11 May 2012)*
523 A small bugfix release.
525 The main issues solved are on the topic of compatibility with newer LVM
528 - fixed parsing of ``lv_attr`` field
529 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
530 the ``--force`` flag is needed
532 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
533 to accept kernel 3.x too (was hardcoded to 2.6.*).
535 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
536 of hook results and that also made display of other errors suboptimal
537 was fixed; the code behaves now like 2.4 and earlier.
539 Another change in 2.5, the cleanup of the OS scripts environment, is too
540 aggressive: it removed even the ``PATH`` variable, which requires the OS
541 scripts to *always* need to export it. Since this is a bit too strict,
542 we now export a minimal PATH, the same that we export for hooks.
544 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
545 integrated into this release.
547 Finally, a few other miscellaneous changes were done (no new features,
548 just small improvements):
550 - Fix ``gnt-group --help`` display
551 - Fix hardcoded Xen kernel path
552 - Fix grow-disk handling of invalid units
553 - Update synopsis for ``gnt-cluster repair-disk-sizes``
554 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
560 *(Released Thu, 12 Apr 2012)*
562 Incompatible/important changes and bugfixes
563 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
565 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
566 resource's ``ip_check`` parameter changed from ``True`` to ``False``
567 to match the underlying LUXI interface.
568 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
569 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
570 not maintain backwards-compatibility as the underlying operation
571 changed in an incompatible way. The RAPI client can talk to old
572 servers, but it needs to be told so as the return value changed.
573 - When creating file-based instances via RAPI, the ``file_driver``
574 parameter no longer defaults to ``loop`` and must be specified.
575 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
577 - Support for the undocumented and deprecated RAPI instance creation
578 request format version 0 has been dropped. Use version 1, supported
579 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
580 - Pyparsing 1.4.6 or above is required, see :doc:`installation
581 documentation <install>`.
582 - The "cluster-verify" hooks are now executed per group by the
583 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
584 if you just run ``gnt-cluster verify``, which generates one opcode per
586 - The environment as passed to the OS scripts is cleared, and thus no
587 environment variables defined in the node daemon's environment will be
588 inherited by the scripts.
589 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
591 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
592 support operations involving multiple node groups.
593 - Offline nodes are ignored when failing over an instance.
594 - Support for KVM version 1.0, which changed the version reporting format
596 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
598 - ``Makefile`` is now compatible with Automake 1.11.2
599 - Includes all bugfixes made in the 2.4 series
604 - The ganeti-htools project has been merged into the ganeti-core source
605 tree and will be built as part of Ganeti (see :doc:`install-quick`).
606 - Implemented support for :doc:`shared storage <design-shared-storage>`.
607 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
608 GPT-style partition tables (requires `parted
609 <http://www.gnu.org/s/parted/>`_).
610 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
611 - Allowed adding tags on instance creation.
612 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
613 - Allow instances to be started in a paused state, enabling the user to
614 see the complete console output on boot using the console.
615 - Added new hypervisor flag to control default reboot behaviour
616 (``reboot_behavior``).
617 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
618 - Improved out-of-band management support:
620 - Added ``gnt-node health`` command reporting the health status of
622 - Added ``gnt-node power`` command to manage power status of nodes.
623 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
625 - Instance migration can fall back to failover if instance is not
627 - Filters can be used when listing nodes, instances, groups and locks;
628 see *ganeti(7)* manpage.
629 - Added post-execution status as variables to :doc:`hooks <hooks>`
631 - Instance tags are exported/imported together with the instance.
632 - When given an explicit job ID, ``gnt-job info`` will work for archived
634 - Jobs can define dependencies on other jobs (not yet supported via
635 RAPI or command line, but used by internal commands and usable via
638 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
641 - Instance failover is now available as a RAPI resource
642 (``/2/instances/[instance_name]/failover``).
643 - ``gnt-instance info`` defaults to static information if primary node
645 - Opcodes have a new ``comment`` attribute.
646 - Added basic SPICE support to KVM hypervisor.
647 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
649 Node group improvements
650 ~~~~~~~~~~~~~~~~~~~~~~~
652 - ``gnt-cluster verify`` has been modified to check groups separately,
653 thereby improving performance.
654 - Node group support has been added to ``gnt-cluster verify-disks``,
655 which now operates per node group.
656 - Watcher has been changed to work better with node groups.
658 - One process and state file per node group.
659 - Slow watcher in one group doesn't block other group's watcher.
661 - Added new command, ``gnt-group evacuate``, to move all instances in a
662 node group to other groups.
663 - Added ``gnt-instance change-group`` to move an instance to another
665 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
666 per-group operations.
667 - Node groups can be tagged.
668 - Some operations switch from an exclusive to a shared lock as soon as
670 - Instance's primary and secondary nodes' groups are now available as
671 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
672 and ``snodes.group.uuid``).
677 - Numerous updates to documentation and manpages.
679 - :doc:`RAPI <rapi>` documentation now has detailed parameter
681 - Some opcode/job results are now also documented, see :doc:`RAPI
684 - A lockset's internal lock is now also visible in lock monitor.
685 - Log messages from job queue workers now contain information about the
686 opcode they're processing.
687 - ``gnt-instance console`` no longer requires the instance lock.
688 - A short delay when waiting for job changes reduces the number of LUXI
689 requests significantly.
690 - DRBD metadata volumes are overwritten with zeros during disk creation.
691 - Out-of-band commands no longer acquire the cluster lock in exclusive
693 - ``devel/upload`` now uses correct permissions for directories.
699 *(Released Fri, 23 Mar 2012)*
701 This was the sixth release candidate of the 2.5 series.
707 *(Released Mon, 9 Jan 2012)*
709 This was the fifth release candidate of the 2.5 series.
715 *(Released Thu, 27 Oct 2011)*
717 This was the fourth release candidate of the 2.5 series.
723 *(Released Wed, 26 Oct 2011)*
725 This was the third release candidate of the 2.5 series.
731 *(Released Tue, 18 Oct 2011)*
733 This was the second release candidate of the 2.5 series.
739 *(Released Tue, 4 Oct 2011)*
741 This was the first release candidate of the 2.5 series.
747 *(Released Wed, 31 Aug 2011)*
749 This was the third beta release of the 2.5 series.
755 *(Released Mon, 22 Aug 2011)*
757 This was the second beta release of the 2.5 series.
763 *(Released Fri, 12 Aug 2011)*
765 This was the first beta release of the 2.5 series.
771 *(Released Thu, 27 Oct 2011)*
773 - Fixed bug when parsing command line parameter values ending in
775 - Fixed assertion error after unclean master shutdown
776 - Disable HTTP client pool for RPC, significantly reducing memory usage
778 - Fixed queue archive creation with wrong permissions
784 *(Released Tue, 23 Aug 2011)*
788 - Fixed documentation for importing with ``--src-dir`` option
789 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
790 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
796 *(Released Fri, 5 Aug 2011)*
798 Many bug-fixes and a few small features:
800 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
801 issues when you tried to add an instance with two MAC addresses in one
803 - KVM: fixed per-instance stored UID value
804 - KVM: configure bridged NICs at migration start
805 - KVM: Fix a bug where instance will not start with never KVM versions
807 - Added OS search path to ``gnt-cluster info``
808 - Fixed an issue with ``file_storage_dir`` where you were forced to
809 provide an absolute path, but the documentation states it is a
810 relative path, the documentation was right
811 - Added a new parameter to instance stop/start called ``--no-remember``
812 that will make the state change to not be remembered
813 - Implemented ``no_remember`` at RAPI level
814 - Improved the documentation
815 - Node evacuation: don't call IAllocator if node is already empty
816 - Fixed bug in DRBD8 replace disks on current nodes
817 - Fixed bug in recreate-disks for DRBD instances
818 - Moved assertion checking locks in ``gnt-instance replace-disks``
819 causing it to abort with not owning the right locks for some situation
820 - Job queue: Fixed potential race condition when cancelling queued jobs
821 - Fixed off-by-one bug in job serial generation
822 - ``gnt-node volumes``: Fix instance names
823 - Fixed aliases in bash completion
824 - Fixed a bug in reopening log files after being sent a SIGHUP
825 - Added a flag to burnin to allow specifying VCPU count
826 - Bugfixes to non-root Ganeti configuration
832 *(Released Thu, 12 May 2011)*
834 Many bug-fixes and a few new small features:
836 - Fixed a bug related to log opening failures
837 - Fixed a bug in instance listing with orphan instances
838 - Fixed a bug which prevented resetting the cluster-level node parameter
839 ``oob_program`` to the default
840 - Many fixes related to the ``cluster-merge`` tool
841 - Fixed a race condition in the lock monitor, which caused failures
842 during (at least) creation of many instances in parallel
843 - Improved output for gnt-job info
844 - Removed the quiet flag on some ssh calls which prevented debugging
846 - Improved the N+1 failure messages in cluster verify by actually
847 showing the memory values (needed and available)
848 - Increased lock attempt timeouts so that when executing long operations
849 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
850 too early and thus prevent the use of the 'fair' mechanism
851 - Changed instance query data (``gnt-instance info``) to not acquire
852 locks unless needed, thus allowing its use on locked instance if only
853 static information is asked for
854 - Improved behaviour with filesystems that do not support rename on an
856 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
857 kept locks on all nodes during the wipe, which is unneeded
858 - Fixed ``gnt-watcher`` handling of errors during hooks execution
859 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
860 10GiB) which caused the wipe to fail right at the end in some cases
861 - Fixed master IP activation when doing master failover with no-voting
862 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
863 the master node itself
864 - Fixed potential data-loss in under disk full conditions, where Ganeti
865 wouldn't check correctly the return code and would consider
866 partially-written files 'correct'
867 - Fixed bug related to multiple VGs and DRBD disk replacing
868 - Added new disk parameter ``metavg`` that allows placement of the meta
869 device for DRBD in a different volume group
870 - Fixed error handling in the node daemon when the system libc doesn't
871 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
872 - Fixed lock release during replace-disks, which kept cluster-wide locks
873 when doing disk replaces with an iallocator script
874 - Added check for missing bridges in cluster verify
875 - Handle EPIPE errors while writing to the terminal better, so that
876 piping the output to e.g. ``less`` doesn't cause a backtrace
877 - Fixed rare case where a ^C during Luxi calls could have been
878 interpreted as server errors, instead of simply terminating
879 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
881 - Added a few more parameters to the KVM hypervisor, allowing a second
882 CDROM, custom disk type for CDROMs and a floppy image
883 - Removed redundant message in instance rename when the name is given
885 - Added option to ``gnt-instance recreate-disks`` to allow creating the
886 disks on new nodes, allowing recreation when the original instance
887 nodes are completely gone
888 - Added option when converting disk templates to DRBD to skip waiting
889 for the resync, in order to make the instance available sooner
890 - Added two new variables to the OS scripts environment (containing the
892 - Made the root_path and optional parameter for the xen-pvm hypervisor,
893 to allow use of ``pvgrub`` as bootloader
894 - Changed the instance memory modifications to only check out-of-memory
895 conditions on memory increases, and turned the secondary node warnings
896 into errors (they can still be overridden via ``--force``)
897 - Fixed the handling of a corner case when the Python installation gets
898 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
899 to execute a command that doesn't exist
900 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
901 node of the instance returned failures during instance shutdown; this
902 adds the option ``--ignore-consistency`` to gnt-instance move
904 And as usual, various improvements to the error messages, documentation
911 *(Released Wed, 09 Mar 2011)*
913 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
914 the RAPI users file if run twice (even with ``--dry-run``).
916 The release fixes that bug (nothing else changed).
922 *(Released Mon, 07 Mar 2011)*
924 Final 2.4.0 release. Just a few small fixes:
926 - Fixed RAPI node evacuate
927 - Fixed the kvm-ifup script
928 - Fixed internal error handling for special job cases
929 - Updated man page to specify the escaping feature for options
935 *(Released Mon, 28 Feb 2011)*
937 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
938 that this feature wiped the disks of the wrong instance, leading to loss
943 - Fixed title of query field containing instance name
944 - Expanded the glossary in the documentation
945 - Fixed one unittest (internal issue)
951 *(Released Mon, 21 Feb 2011)*
953 A number of bug fixes plus just a couple functionality changes.
955 On the user-visible side, the ``gnt-* list`` command output has changed
956 with respect to "special" field states. The current rc1 style of display
957 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
958 the default output mode special fields states are displayed as follows:
960 - Offline resource: ``*``
961 - Unavailable/not applicable: ``-``
962 - Data missing (RPC failure): ``?``
963 - Unknown field: ``??``
965 Another user-visible change is the addition of ``--force-join`` to
970 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
971 - Fixed regression in RAPI/instance reinstall where all parameters were
972 required (instead of optional)
973 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
974 - Fixed iallocator usage (offline nodes were not considered offline)
975 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
976 - Fixed hypervisor and OS parameter validation with respect to
978 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
980 - Fixed ``tools/listrunner`` with respect to agent-based usage
986 *(Released Fri, 4 Feb 2011)*
988 Many changes and fixes since the beta1 release. While there were some
989 internal changes, the code has been mostly stabilised for the RC
992 Note: the dumb allocator was removed in this release, as it was not kept
993 up-to-date with the IAllocator protocol changes. It is recommended to
994 use the ``hail`` command from the ganeti-htools package.
996 Note: the 2.4 and up versions of Ganeti are not compatible with the
997 0.2.x branch of ganeti-htools. You need to upgrade to
998 ganeti-htools-0.3.0 (or later).
1000 Regressions fixed from 2.3
1001 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1003 - Fixed the ``gnt-cluster verify-disks`` command
1004 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1006 - Fixed disk adoption breakage
1007 - Fixed wrong headers in instance listing for field aliases
1012 - Fixed corner case in KVM handling of NICs
1013 - Fixed many cases of wrong handling of non-vm_capable nodes
1014 - Fixed a bug where a missing instance symlink was not possible to
1015 recreate with any ``gnt-*`` command (now ``gnt-instance
1016 activate-disks`` does it)
1017 - Fixed the volume group name as reported by ``gnt-cluster
1019 - Increased timeouts for the import-export code, hopefully leading to
1020 fewer aborts due network or instance timeouts
1021 - Fixed bug in ``gnt-node list-storage``
1022 - Fixed bug where not all daemons were started on cluster
1023 initialisation, but only at the first watcher run
1024 - Fixed many bugs in the OOB implementation
1025 - Fixed watcher behaviour in presence of instances with offline
1027 - Fixed instance list output for instances running on the wrong node
1028 - a few fixes to the cluster-merge tool, but it still cannot merge
1029 multi-node groups (currently it is not recommended to use this tool)
1035 - Improved network configuration for the KVM hypervisor
1036 - Added e1000 as a supported NIC for Xen-HVM
1037 - Improved the lvmstrap tool to also be able to use partitions, as
1038 opposed to full disks
1039 - Improved speed of disk wiping (the cluster parameter
1040 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1041 of instance creations
1042 - Added documentation for the OS parameters
1043 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1044 hypervisor is not responding
1045 - Added display of blacklisted and hidden OS information in
1046 ``gnt-cluster info``
1047 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1048 NIC and node parameters, which might create problems with currently
1049 invalid (but undetected) configuration files, but prevents validation
1050 failures when unrelated parameters are modified
1051 - Changed cluster initialisation to wait for the master daemon to become
1053 - Expanded the RAPI interface:
1055 - Added config redistribution resource
1056 - Added activation/deactivation of instance disks
1057 - Added export of console information
1059 - Implemented log file reopening on SIGHUP, which allows using
1060 logrotate(8) for the Ganeti log files
1061 - Added a basic OOB helper script as an example
1067 *(Released Fri, 14 Jan 2011)*
1072 - Fixed timezone issues when formatting timestamps
1073 - Added support for node groups, available via ``gnt-group`` and other
1075 - Added out-of-band framework and management, see :doc:`design
1076 document <design-oob>`
1077 - Removed support for roman numbers from ``gnt-node list`` and
1078 ``gnt-instance list``.
1079 - Allowed modification of master network interface via ``gnt-cluster
1080 modify --master-netdev``
1081 - Accept offline secondaries while shutting down instance disks
1082 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1083 - Added support for multiple LVM volume groups
1084 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1086 - Added commands to list available fields:
1088 - ``gnt-node list-fields``
1089 - ``gnt-group list-fields``
1090 - ``gnt-instance list-fields``
1092 - Updated documentation and man pages
1097 - Moved ``rapi_users`` file into separate directory, now named
1098 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1100 - Added new tool for running commands on many machines,
1101 ``tools/ganeti-listrunner``
1102 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1103 improving the ``gnt-instance console`` output
1104 - Allowed customisation of disk index separator at ``configure`` time
1105 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1106 - Added support for non-partitioned md disks in ``lvmstrap``
1107 - Added script to gracefully power off KVM instances
1108 - Split ``utils`` module into smaller parts
1109 - Changed query operations to return more detailed information, e.g.
1110 whether an information is unavailable due to an offline node. To use
1111 this new functionality, the LUXI call ``Query`` must be used. Field
1112 information is now stored by the master daemon and can be retrieved
1113 using ``QueryFields``. Instances, nodes and groups can also be queried
1114 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1115 exposed via RAPI). The following commands make use of this
1116 infrastructure change:
1118 - ``gnt-group list``
1119 - ``gnt-group list-fields``
1121 - ``gnt-node list-fields``
1122 - ``gnt-instance list``
1123 - ``gnt-instance list-fields``
1124 - ``gnt-debug locks``
1129 - New RAPI resources (see :doc:`rapi`):
1133 - ``/2/groups/[group_name]``
1134 - ``/2/groups/[group_name]/assign-nodes``
1135 - ``/2/groups/[group_name]/modify``
1136 - ``/2/groups/[group_name]/rename``
1137 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1141 - Implemented ``no_install`` for instance creation
1142 - Implemented OS parameters for instance reinstallation, allowing
1143 use of special settings on reinstallation (e.g. for preserving data)
1148 - Added IPv6 support in import/export
1149 - Pause DRBD synchronization while wiping disks on instance creation
1150 - Updated unittests and QA scripts
1151 - Improved network parameters passed to KVM
1152 - Converted man pages from docbook to reStructuredText
1158 *(Released Mon, 20 Dec 2010)*
1160 Released version 2.3.1~rc1 without any changes.
1166 *(Released Wed, 1 Dec 2010)*
1168 - impexpd: Disable OpenSSL compression in socat if possible (backport
1169 from master, commit e90739d625b, see :doc:`installation guide
1170 <install-quick>` for details)
1171 - Changed unittest coverage report to exclude test scripts
1172 - Added script to check version format
1178 *(Released Wed, 1 Dec 2010)*
1180 Released version 2.3.0~rc1 without any changes.
1186 *(Released Fri, 19 Nov 2010)*
1188 A number of bugfixes and documentation updates:
1190 - Update ganeti-os-interface documentation
1191 - Fixed a bug related to duplicate MACs or similar items which should be
1193 - Fix breakage in OS state modify
1194 - Reinstall instance: disallow offline secondaries (fixes bug related to
1195 OS changing but reinstall failing)
1196 - plus all the other fixes between 2.2.1 and 2.2.2
1202 *(Released Tue, 2 Nov 2010)*
1204 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1205 - Fixed master failover race with watcher
1206 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1208 - Accept previously stopped instance for export with instance removal
1209 - Simplify and extend the environment variables for instance OS scripts
1210 - Added new node flags, ``master_capable`` and ``vm_capable``
1211 - Added optional instance disk wiping prior during allocation. This is a
1212 cluster-wide option and can be set/modified using
1213 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1214 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1215 :doc:`install-quick`
1216 - Added a new watcher option (``--ignore-pause``)
1217 - Added option to ignore offline node on instance start/stop
1218 (``--ignore-offline``)
1219 - Allow overriding OS parameters with ``gnt-instance reinstall``
1220 - Added ability to change node's secondary IP address using ``gnt-node
1222 - Implemented privilege separation for all daemons except
1223 ``ganeti-noded``, see ``configure`` options
1224 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1226 - Implemented job priorities (see ``ganeti(7)`` manpage)
1227 - Ignore failures while shutting down instances during failover from
1229 - Exit daemon's bootstrap process only once daemon is ready
1230 - Export more information via ``LUInstanceQuery``/remote API
1231 - Improved documentation, QA and unittests
1232 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1233 restart if the file was created or changed
1234 - Added LUXI protocol version sent with each request and response,
1235 allowing detection of server/client mismatches
1236 - Moved the Python scripts among gnt-* and ganeti-* into modules
1237 - Moved all code related to setting up SSH to an external script,
1239 - Infrastructure changes for node group support in future versions
1245 *(Released Fri, 19 Nov 2010)*
1247 A few small bugs fixed, and some improvements to the build system:
1249 - Fix documentation regarding conversion to drbd
1250 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1252 - Fix error handling in node modify with multiple changes
1253 - Allow remote imports without checked names
1259 *(Released Tue, 19 Oct 2010)*
1261 - Disable SSL session ID cache in RPC client
1267 *(Released Thu, 14 Oct 2010)*
1269 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1270 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1272 - Fix problems with interaction between Curl and Python's HTTP server,
1273 resulting in increased speed in many RPC calls
1274 - Improve our release script to prevent breakage with older aclocal and
1281 *(Released Thu, 7 Oct 2010)*
1283 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1284 value retrieved from master
1285 - Added support for blacklisted or hidden OS definitions
1286 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1287 - Added support for -mem-path in KVM hypervisor abstraction layer
1288 - Allow overriding instance parameters in tool for inter-cluster
1289 instance moves (``tools/move-instance``)
1290 - Improved opcode summaries (e.g. in ``gnt-job list``)
1291 - Improve consistency of OS listing by sorting it
1292 - Documentation updates
1298 *(Released Fri, 8 Oct 2010)*
1300 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1306 *(Released Mon, 4 Oct 2010)*
1308 - Fixed regression in ``gnt-instance rename``
1314 *(Released Wed, 22 Sep 2010)*
1316 - Fixed OS_VARIANT variable for OS scripts
1317 - Fixed cluster tag operations via RAPI
1318 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1319 - Disabled RAPI CA checks in watcher
1325 *(Released Mon, 23 Aug 2010)*
1327 - Support DRBD versions of the format "a.b.c.d"
1329 - Re-introduce support for usage from multiple threads in RAPI client
1330 - Instance renames and modify via RAPI
1331 - Work around race condition between processing and archival in job
1333 - Mark opcodes following failed one as failed, too
1334 - Job field ``lock_status`` was removed due to difficulties making it
1335 work with the changed job queue in Ganeti 2.2; a better way to monitor
1336 locks is expected for a later 2.2.x release
1337 - Fixed dry-run behaviour with many commands
1338 - Support ``ssh-agent`` again when adding nodes
1339 - Many additional bugfixes
1345 *(Released Fri, 30 Jul 2010)*
1347 Important change: the internal RPC mechanism between Ganeti nodes has
1348 changed from using a home-grown http library (based on the Python base
1349 libraries) to use the PycURL library. This requires that PycURL is
1350 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1351 against GnuTLS by default. cURL's support for GnuTLS had known issues
1352 before cURL 7.21.0 and we recommend using the latest cURL release or
1353 linking against OpenSSL. Most other distributions already link PycURL
1354 and cURL against OpenSSL. The command::
1356 python -c 'import pycurl; print pycurl.version'
1358 can be used to determine the libraries PycURL and cURL are linked
1361 Other significant changes:
1363 - Rewrote much of the internals of the job queue, in order to achieve
1364 better parallelism; this decouples job query operations from the job
1365 processing, and it should allow much nicer behaviour of the master
1366 daemon under load, and it also has uncovered some long-standing bugs
1367 related to the job serialisation (now fixed)
1368 - Added a default iallocator setting to the cluster parameters,
1369 eliminating the need to always pass nodes or an iallocator for
1370 operations that require selection of new node(s)
1371 - Added experimental support for the LXC virtualization method
1372 - Added support for OS parameters, which allows the installation of
1373 instances to pass parameter to OS scripts in order to customise the
1375 - Added a hypervisor parameter controlling the migration type (live or
1376 non-live), since hypervisors have various levels of reliability; this
1377 has renamed the 'live' parameter to 'mode'
1378 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1379 logical volumes, meaning that cluster verify will ignore them and not
1380 flag their presence as errors
1381 - The watcher will now reset the error count for failed instances after
1382 8 hours, thus allowing self-healing if the problem that caused the
1383 instances to be down/fail to start has cleared in the meantime
1384 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1385 check for, and warn, if the drbd module parameter ``usermode_helper``
1386 is not consistent with the cluster-wide setting; this is needed to
1387 make diagnose easier of failed drbd creations
1388 - Started adding base IPv6 support, but this is not yet
1389 enabled/available for use
1390 - Rename operations (cluster, instance) will now return the new name,
1391 which is especially useful if a short name was passed in
1392 - Added support for instance migration in RAPI
1393 - Added a tool to pre-configure nodes for the SSH setup, before joining
1394 them to the cluster; this will allow in the future a simplified model
1395 for node joining (but not yet fully enabled in 2.2); this needs the
1396 paramiko python library
1397 - Fixed handling of name-resolving errors
1398 - Fixed consistency of job results on the error path
1399 - Fixed master-failover race condition when executed multiple times in
1401 - Fixed many bugs related to the job queue (mostly introduced during the
1402 2.2 development cycle, so not all are impacting 2.1)
1403 - Fixed instance migration with missing disk symlinks
1404 - Fixed handling of unknown jobs in ``gnt-job archive``
1405 - And many other small fixes/improvements
1409 - Enhanced both the unittest and the QA coverage
1410 - Switched the opcode validation to a generic model, and extended the
1411 validation to all opcode parameters
1412 - Changed more parts of the code that write shell scripts to use the
1414 - Switched the master daemon to use the asyncore library for the Luxi
1421 *(Released Thu, 17 Jun 2010)*
1423 - Added tool (``move-instance``) and infrastructure to move instances
1424 between separate clusters (see :doc:`separate documentation
1425 <move-instance>` and :doc:`design document <design-2.2>`)
1426 - Added per-request RPC timeout
1427 - RAPI now requires a Content-Type header for requests with a body (e.g.
1428 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1429 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1430 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1432 - Implemented initial support for running Ganeti daemons as separate
1433 users, see configure-time flags ``--with-user-prefix`` and
1434 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1436 - Instances can be removed after export (``gnt-backup export
1437 --remove-instance``)
1438 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1439 key (instead of 1024 bit)
1440 - Added new cluster configuration file for cluster domain secret
1441 - Import/export now use SSL instead of SSH
1442 - Added support for showing estimated time when exporting an instance,
1443 see the ``ganeti-os-interface(7)`` manpage and look for
1450 *(Released Tue, 16 Nov 2010)*
1452 Some more bugfixes. Unless critical bugs occur, this will be the last
1455 - Fix case of MAC special-values
1456 - Fix mac checker regex
1457 - backend: Fix typo causing "out of range" error
1458 - Add missing --units in gnt-instance list man page
1464 *(Released Tue, 24 Aug 2010)*
1467 - Don't ignore secondary node silently on non-mirrored disk templates
1469 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1470 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1471 - Properly document the usb_mouse parameter
1472 - Fix path in ganeti-rapi(8) (issue 116)
1473 - Adjust error message when the ganeti user's .ssh directory is
1475 - Add same-node-check when changing the disk template to drbd
1481 *(Released Fri, 16 Jul 2010)*
1484 - Add an option to only select some reboot types during qa/burnin.
1485 (on some hypervisors consequent reboots are not supported)
1486 - Fix infrequent race condition in master failover. Sometimes the old
1487 master ip address would be still detected as up for a short time
1488 after it was removed, causing failover to fail.
1489 - Decrease mlockall warnings when the ctypes module is missing. On
1490 Python 2.4 we support running even if no ctypes module is installed,
1491 but we were too verbose about this issue.
1492 - Fix building on old distributions, on which man doesn't have a
1494 - Fix RAPI not to ignore the MAC address on instance creation
1495 - Implement the old instance creation format in the RAPI client.
1501 *(Released Thu, 01 Jul 2010)*
1503 A small bugfix release:
1504 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1505 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1506 a non-existing option
1507 - Fix instance create: the --force-variant option was ignored
1508 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1509 - Fix modify node storage with non-FQDN arguments
1510 - Fix RAPI client to authenticate under Python 2.6 when used
1511 for more than 5 requests needing authentication
1512 - Fix gnt-instance modify -t (storage) giving a wrong error message
1513 when converting a non-shutdown drbd instance to plain
1519 *(Released Fri, 18 Jun 2010)*
1521 A small bugfix release:
1523 - Fix live migration of KVM instances started with older Ganeti
1524 versions which had fewer hypervisor parameters
1525 - Fix gnt-instance grow-disk on down instances
1526 - Fix an error-reporting bug during instance migration
1527 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1528 silently ignoring broken ones
1529 - Fix an RPC error reporting bug affecting, for example, RAPI client
1531 - Fix bug triggered by different API version os-es on different nodes
1532 - Fix a bug in instance startup with custom hvparams: OS level
1533 parameters would fail to be applied.
1534 - Fix the RAPI client under Python 2.6 (but more work is needed to
1535 make it work completely well with OpenSSL)
1536 - Fix handling of errors when resolving names from DNS
1542 *(Released Thu, 3 Jun 2010)*
1544 A medium sized development cycle. Some new features, and some
1545 fixes/small improvements/cleanups.
1547 Significant features
1548 ~~~~~~~~~~~~~~~~~~~~
1550 The node deamon now tries to mlock itself into memory, unless the
1551 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1552 its logs, and falls back to console logging. This allows emergency
1553 features such as ``gnt-node powercycle`` to work even in the event of a
1554 broken node disk (tested offlining the disk hosting the node's
1555 filesystem and dropping its memory caches; don't try this at home)
1557 KVM: add vhost-net acceleration support. It can be tested with a new
1558 enough version of the kernel and of qemu-kvm.
1560 KVM: Add instance chrooting feature. If you use privilege dropping for
1561 your VMs you can also now force them to chroot to an empty directory,
1562 before starting the emulated guest.
1564 KVM: Add maximum migration bandwith and maximum downtime tweaking
1565 support (requires a new-enough version of qemu-kvm).
1567 Cluster verify will now warn if the master node doesn't have the master
1568 ip configured on it.
1570 Add a new (incompatible) instance creation request format to RAPI which
1571 supports all parameters (previously only a subset was supported, and it
1572 wasn't possible to extend the old format to accomodate all the new
1573 features. The old format is still supported, and a client can check for
1574 this feature, before using it, by checking for its presence in the
1575 ``features`` RAPI resource.
1577 Now with ancient latin support. Try it passing the ``--roman`` option to
1578 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1579 (requires the python-roman module to be installed, in order to work).
1584 As usual many internal code refactorings, documentation updates, and
1587 - Lots of improvements and cleanups to the experimental Remote API
1588 (RAPI) client library.
1589 - A new unit test suite for the core daemon libraries.
1590 - A fix to creating missing directories makes sure the umask is not
1591 applied anymore. This enforces the same directory permissions
1593 - Better handling terminating daemons with ctrl+c (used when running
1594 them in debugging mode).
1595 - Fix a race condition in live migrating a KVM instance, when stat()
1596 on the old proc status file returned EINVAL, which is an unexpected
1598 - Fixed manpage checking with newer man and utf-8 charachters. But now
1599 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1605 *(Released Fri, 7 May 2010)*
1607 Fix a bug which prevented untagged KVM instances from starting.
1613 *(Released Fri, 7 May 2010)*
1615 Another release with a long development cycle, during which many
1616 different features were added.
1618 Significant features
1619 ~~~~~~~~~~~~~~~~~~~~
1621 The KVM hypervisor now can run the individual instances as non-root, to
1622 reduce the impact of a VM being hijacked due to bugs in the
1623 hypervisor. It is possible to run all instances as a single (non-root)
1624 user, to manually specify a user for each instance, or to dynamically
1625 allocate a user out of a cluster-wide pool to each instance, with the
1626 guarantee that no two instances will run under the same user ID on any
1629 An experimental RAPI client library, that can be used standalone
1630 (without the other Ganeti libraries), is provided in the source tree as
1631 ``lib/rapi/client.py``. Note this client might change its interface in
1632 the future, as we iterate on its capabilities.
1634 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1635 replace the cluster's certificates and crypto keys. This might help in
1636 case they have been compromised, or have simply expired.
1638 A new disk option for instance creation has been added that allows one
1639 to "adopt" currently existing logical volumes, with data
1640 preservation. This should allow easier migration to Ganeti from
1641 unmanaged (or managed via other software) instances.
1643 Another disk improvement is the possibility to convert between redundant
1644 (DRBD) and plain (LVM) disk configuration for an instance. This should
1645 allow better scalability (starting with one node and growing the
1646 cluster, or shrinking a two-node cluster to one node).
1648 A new feature that could help with automated node failovers has been
1649 implemented: if a node sees itself as offline (by querying the master
1650 candidates), it will try to shutdown (hard) all instances and any active
1651 DRBD devices. This reduces the risk of duplicate instances if an
1652 external script automatically failovers the instances on such nodes. To
1653 enable this, the cluster parameter ``maintain_node_health`` should be
1654 enabled; in the future this option (per the name) will enable other
1655 automatic maintenance features.
1657 Instance export/import now will reuse the original instance
1658 specifications for all parameters; that means exporting an instance,
1659 deleting it and the importing it back should give an almost identical
1660 instance. Note that the default import behaviour has changed from
1661 before, where it created only one NIC; now it recreates the original
1664 Cluster verify has added a few new checks: SSL certificates validity,
1665 /etc/hosts consistency across the cluster, etc.
1670 As usual, many internal changes were done, documentation fixes,
1673 - Fixed cluster initialization with disabled cluster storage (regression
1674 introduced in 2.1.1)
1675 - File-based storage supports growing the disks
1676 - Fixed behaviour of node role changes
1677 - Fixed cluster verify for some corner cases, plus a general rewrite of
1678 cluster verify to allow future extension with more checks
1679 - Fixed log spamming by watcher and node daemon (regression introduced
1681 - Fixed possible validation issues when changing the list of enabled
1683 - Fixed cleanup of /etc/hosts during node removal
1684 - Fixed RAPI response for invalid methods
1685 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1686 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1687 from ide disks, etc.)
1688 - Allow OS changes without re-installation (to record a changed OS
1689 outside of Ganeti, or to allow OS renames)
1690 - Allow instance creation without OS installation (useful for example if
1691 the OS will be installed manually, or restored from a backup not in
1693 - Implemented option to make cluster ``copyfile`` use the replication
1695 - Added list of enabled hypervisors to ssconf (possibly useful for
1697 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1699 - A partial form of node re-IP is possible via node readd, which now
1700 allows changed node primary IP
1701 - Command line utilities now show an informational message if the job is
1703 - The logs of the master daemon now show the PID/UID/GID of the
1710 *(Released Fri, 12 Mar 2010)*
1712 During the 2.1.0 long release candidate cycle, a lot of improvements and
1713 changes have accumulated with were released later as 2.1.1.
1718 The node evacuate command (``gnt-node evacuate``) was significantly
1719 rewritten, and as such the IAllocator protocol was changed - a new
1720 request type has been added. This unfortunate change during a stable
1721 series is designed to improve performance of node evacuations; on
1722 clusters with more than about five nodes and which are well-balanced,
1723 evacuation should proceed in parallel for all instances of the node
1724 being evacuated. As such, any existing IAllocator scripts need to be
1725 updated, otherwise the above command will fail due to the unknown
1726 request. The provided "dumb" allocator has not been updated; but the
1727 ganeti-htools package supports the new protocol since version 0.2.4.
1729 Another important change is increased validation of node and instance
1730 names. This might create problems in special cases, if invalid host
1731 names are being used.
1733 Also, a new layer of hypervisor parameters has been added, that sits at
1734 OS level between the cluster defaults and the instance ones. This allows
1735 customisation of virtualization parameters depending on the installed
1736 OS. For example instances with OS 'X' may have a different KVM kernel
1737 (or any other parameter) than the cluster defaults. This is intended to
1738 help managing a multiple OSes on the same cluster, without manual
1739 modification of each instance's parameters.
1741 A tool for merging clusters, ``cluster-merge``, has been added in the
1742 tools sub-directory.
1747 - Improved the int/float conversions that should make the code more
1748 robust in face of errors from the node daemons
1749 - Fixed the remove node code in case of internal configuration errors
1750 - Fixed the node daemon behaviour in face of inconsistent queue
1751 directory (e.g. read-only file-system where we can't open the files
1753 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1754 now it either aborts cleanly or, if given the new "auto_promote"
1755 parameter, will automatically promote other nodes as needed
1756 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1757 completely prevent Ganeti from working
1758 - Fixed bug for instance export when not all disks were successfully
1760 - Fixed behaviour of node add when the new node is slow in starting up
1762 - Fixed handling of signals in the LUXI client, which should improve
1763 behaviour of command-line scripts
1764 - Added checks for invalid node/instance names in the configuration (now
1765 flagged during cluster verify)
1766 - Fixed watcher behaviour for disk activation errors
1767 - Fixed two potentially endless loops in http library, which led to the
1768 RAPI daemon hanging and consuming 100% CPU in some cases
1769 - Fixed bug in RAPI daemon related to hashed passwords
1770 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1772 - Enhanced compatibility with non-Debian OSes, but not using absolute
1773 path in some commands and allowing customisation of the ssh
1774 configuration directory
1775 - Fixed possible future issue with new Python versions by abiding to the
1776 proper use of ``__slots__`` attribute on classes
1777 - Added checks that should prevent directory traversal attacks
1778 - Many documentation fixes based on feedback from users
1783 - Added an "early_release" more for instance replace disks and node
1784 evacuate, where we release locks earlier and thus allow higher
1785 parallelism within the cluster
1786 - Added watcher hooks, intended to allow the watcher to restart other
1787 daemons (e.g. from the ganeti-nbma project), but they can be used of
1788 course for any other purpose
1789 - Added a compile-time disable for DRBD barriers, to increase
1790 performance if the administrator trusts the power supply or the
1791 storage system to not lose writes
1792 - Added the option of using syslog for logging instead of, or in
1793 addition to, Ganeti's own log files
1794 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1795 can indeed boot from a paravirtual NIC
1796 - Added a generic debug level for many operations; while this is not
1797 used widely yet, it allows one to pass the debug value all the way to
1799 - Enhanced the hooks environment for instance moves (failovers,
1800 migrations) where the primary/secondary nodes changed during the
1801 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1802 - Enhanced data validations for many user-supplied values; one important
1803 item is the restrictions imposed on instance and node names, which
1804 might reject some (invalid) host names
1805 - Add a configure-time option to disable file-based storage, if it's not
1806 needed; this allows greater security separation between the master
1807 node and the other nodes from the point of view of the inter-node RPC
1809 - Added user notification in interactive tools if job is waiting in the
1810 job queue or trying to acquire locks
1811 - Added log messages when a job is waiting for locks
1812 - Added filtering by node tags in instance operations which admit
1813 multiple instances (start, stop, reboot, reinstall)
1814 - Added a new tool for cluster mergers, ``cluster-merge``
1815 - Parameters from command line which are of the form ``a=b,c=d`` can now
1816 use backslash escapes to pass in values which contain commas,
1817 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1819 - For KVM, the instance name is the first parameter passed to KVM, so
1820 that it's more visible in the process list
1826 *(Released Tue, 2 Mar 2010)*
1828 Ganeti 2.1 brings many improvements with it. Major changes:
1830 - Added infrastructure to ease automated disk repairs
1831 - Added new daemon to export configuration data in a cheaper way than
1832 using the remote API
1833 - Instance NICs can now be routed instead of being associated with a
1835 - Improved job locking logic to reduce impact of jobs acquiring multiple
1836 locks waiting for other long-running jobs
1838 In-depth implementation details can be found in the Ganeti 2.1 design
1844 - Added chroot hypervisor
1845 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1847 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1848 ``bootloader_path`` and ``bootloader_args``)
1849 - Added the ``use_localtime`` option for the xen-hvm and kvm
1850 hypervisors, and the default value for this has changed to false (in
1851 2.0 xen-hvm always enabled it)
1852 - Added luxi call to submit multiple jobs in one go
1853 - Added cluster initialization option to not modify ``/etc/hosts``
1855 - Added network interface parameters
1856 - Added dry run mode to some LUs
1857 - Added RAPI resources:
1859 - ``/2/instances/[instance_name]/info``
1860 - ``/2/instances/[instance_name]/replace-disks``
1861 - ``/2/nodes/[node_name]/evacuate``
1862 - ``/2/nodes/[node_name]/migrate``
1863 - ``/2/nodes/[node_name]/role``
1864 - ``/2/nodes/[node_name]/storage``
1865 - ``/2/nodes/[node_name]/storage/modify``
1866 - ``/2/nodes/[node_name]/storage/repair``
1868 - Added OpCodes to evacuate or migrate all instances on a node
1869 - Added new command to list storage elements on nodes (``gnt-node
1870 list-storage``) and modify them (``gnt-node modify-storage``)
1871 - Added new ssconf files with master candidate IP address
1872 (``ssconf_master_candidates_ips``), node primary IP address
1873 (``ssconf_node_primary_ips``) and node secondary IP address
1874 (``ssconf_node_secondary_ips``)
1875 - Added ``ganeti-confd`` and a client library to query the Ganeti
1876 configuration via UDP
1877 - Added ability to run hooks after cluster initialization and before
1879 - Added automatic mode for disk replace (``gnt-instance replace-disks
1881 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1882 after catastrophic data-loss
1883 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1885 - Added ``gnt-instance move`` command to move instances
1886 - Added ``gnt-cluster watcher`` command to control watcher
1887 - Added ``gnt-node powercycle`` command to powercycle nodes
1888 - Added new job status field ``lock_status``
1889 - Added parseable error codes to cluster verification (``gnt-cluster
1890 verify --error-codes``) and made output less verbose (use
1891 ``--verbose`` to restore previous behaviour)
1892 - Added UUIDs to the main config entities (cluster, nodes, instances)
1893 - Added support for OS variants
1894 - Added support for hashed passwords in the Ganeti remote API users file
1896 - Added option to specify maximum timeout on instance shutdown
1897 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
1898 - Added new helper script to start and stop Ganeti daemons
1899 (``daemon-util``), with the intent to reduce the work necessary to
1900 adjust Ganeti for non-Debian distributions and to start/stop daemons
1902 - Added more unittests
1903 - Fixed critical bug in ganeti-masterd startup
1904 - Removed the configure-time ``kvm-migration-port`` parameter, this is
1905 now customisable at the cluster level for both the KVM and Xen
1906 hypervisors using the new ``migration_port`` parameter
1907 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
1908 reinstalling an instance
1909 - Allowed ``@`` in tag names
1910 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
1911 - Many documentation updates
1912 - Distribute hypervisor files on ``gnt-cluster redist-conf``
1913 - ``gnt-instance reinstall`` can now reinstall multiple instances
1914 - Updated many command line parameters
1915 - Introduced new OS API version 15
1916 - No longer support a default hypervisor
1917 - Treat virtual LVs as inexistent
1918 - Improved job locking logic to reduce lock contention
1919 - Match instance and node names case insensitively
1920 - Reimplemented bash completion script to be more complete
1927 *(Released Thu, 4 Feb 2010)*
1929 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
1930 - Fix a string formatting bug
1931 - Improve safety of the code in some error paths
1932 - Improve data validation in the master of values returned from nodes
1938 *(Released Thu, 17 Dec 2009)*
1940 - Fix security issue due to missing validation of iallocator names; this
1941 allows local and remote execution of arbitrary executables
1942 - Fix failure of gnt-node list during instance removal
1943 - Ship the RAPI documentation in the archive
1949 *(Released Wed, 30 Sep 2009)*
1951 - Fixed many wrong messages
1952 - Fixed a few bugs related to the locking library
1953 - Fixed MAC checking at instance creation time
1954 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
1955 - Fixed a few issues related to signal handling in both daemons and
1957 - Fixed the example startup script provided
1958 - Fixed insserv dependencies in the example startup script (patch from
1960 - Fixed handling of drained nodes in the iallocator framework
1961 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
1963 - Fixed error related to invalid job IDs in job polling
1964 - Fixed job/opcode persistence on unclean master shutdown
1965 - Fixed handling of partial job processing after unclean master
1967 - Fixed error reporting from LUs, previously all errors were converted
1968 into execution errors
1969 - Fixed error reporting from burnin
1970 - Decreased significantly the memory usage of the job queue
1971 - Optimised slightly multi-job submission
1972 - Optimised slightly opcode loading
1973 - Backported the multi-job submit framework from the development
1974 branch; multi-instance start and stop should be faster
1975 - Added script to clean archived jobs after 21 days; this will reduce
1976 the size of the queue directory
1977 - Added some extra checks in disk size tracking
1978 - Added an example ethers hook script
1979 - Added a cluster parameter that prevents Ganeti from modifying of
1981 - Added more node information to RAPI responses
1982 - Added a ``gnt-job watch`` command that allows following the ouput of a
1984 - Added a bind-address option to ganeti-rapi
1985 - Added more checks to the configuration verify
1986 - Enhanced the burnin script such that some operations can be retried
1988 - Converted instance reinstall to multi-instance model
1994 *(Released Fri, 7 Aug 2009)*
1996 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
1997 to allow using the pre-2.0.2 behaviour in activation, if any existing
1998 instances have mismatched disk sizes in the configuration
1999 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2000 any configuration mismatches for disk sizes
2001 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2002 failover to work on two-node clusters
2003 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2005 - Fixed detection of OS script errors in ``gnt-backup export``
2006 - Fixed exit code of ``gnt-backup export``
2012 *(Released Fri, 17 Jul 2009)*
2014 - Added experimental support for stripped logical volumes; this should
2015 enhance performance but comes with a higher complexity in the block
2016 device handling; stripping is only enabled when passing
2017 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2018 affected even in the non-stripped mode
2019 - Improved resiliency against transient failures at the end of DRBD
2020 resyncs, and in general of DRBD resync checks
2021 - Fixed a couple of issues with exports and snapshot errors
2022 - Fixed a couple of issues in instance listing
2023 - Added display of the disk size in ``gnt-instance info``
2024 - Fixed checking for valid OSes in instance creation
2025 - Fixed handling of the "vcpus" parameter in instance listing and in
2026 general of invalid parameters
2027 - Fixed http server library, and thus RAPI, to handle invalid
2028 username/password combinations correctly; this means that now they
2029 report unauthorized for queries too, not only for modifications,
2030 allowing earlier detect of configuration problems
2031 - Added a new "role" node list field, equivalent to the master/master
2032 candidate/drained/offline flags combinations
2033 - Fixed cluster modify and changes of candidate pool size
2034 - Fixed cluster verify error messages for wrong files on regular nodes
2035 - Fixed a couple of issues with node demotion from master candidate role
2036 - Fixed node readd issues
2037 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2038 - Added a new ``--no-voting`` option for masterfailover to fix failover
2039 on two-nodes clusters when the former master node is unreachable
2040 - Added instance reinstall over RAPI
2046 *(Released Tue, 16 Jun 2009)*
2048 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2049 allow re-adding the start in single-user option (regression from 1.2)
2050 - the watcher writes the instance status to a file, to allow monitoring
2051 to report the instance status (from the master) based on cached
2052 results of the watcher's queries; while this can get stale if the
2053 watcher is being locked due to other work on the cluster, this is
2054 still an improvement
2055 - the watcher now also restarts the node daemon and the rapi daemon if
2057 - fixed the watcher to handle full and drained queue cases
2058 - hooks export more instance data in the environment, which helps if
2059 hook scripts need to take action based on the instance's properties
2060 (no longer need to query back into ganeti)
2061 - instance failovers when the instance is stopped do not check for free
2062 RAM, so that failing over a stopped instance is possible in low memory
2064 - rapi uses queries for tags instead of jobs (for less job traffic), and
2065 for cluster tags it won't talk to masterd at all but read them from
2067 - a couple of error handling fixes in RAPI
2068 - drbd handling: improved the error handling of inconsistent disks after
2069 resync to reduce the frequency of "there are some degraded disks for
2070 this instance" messages
2071 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2072 error handling path called a wrong function name)
2078 *(Released Wed, 27 May 2009)*
2080 - no changes from rc5
2086 *(Released Wed, 20 May 2009)*
2088 - fix a couple of bugs (validation, argument checks)
2089 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2090 - some small improvements to RAPI and IAllocator
2091 - make watcher automatically start the master daemon if down
2097 *(Released Mon, 27 Apr 2009)*
2099 - change the OS list to not require locks; this helps with big clusters
2100 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2101 volume group is broken
2102 - ``gnt-instance info``, without any arguments, doesn't run for all
2103 instances anymore; either pass ``--all`` or pass the desired
2104 instances; this helps against mistakes on big clusters where listing
2105 the information for all instances takes a long time
2106 - miscellaneous doc and man pages fixes
2112 *(Released Wed, 8 Apr 2009)*
2114 - Change the internal locking model of some ``gnt-node`` commands, in
2115 order to reduce contention (and blocking of master daemon) when
2116 batching many creation/reinstall jobs
2117 - Fixes to Xen soft reboot
2118 - No longer build documentation at build time, instead distribute it in
2119 the archive, in order to reduce the need for the whole docbook/rst
2126 *(Released Fri, 27 Mar 2009)*
2128 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2129 - Fix watcher startup sequence, improves the behaviour of busy clusters
2130 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2131 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2132 bind address and other places
2133 - Some documentation fixes and updates
2139 *(Released Mon, 2 Mar 2009)*
2141 - More documentation updates, now all docs should be more-or-less
2143 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2145 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2147 - Fix an issue related to $libdir/run/ganeti and cluster creation
2153 *(Released Thu, 19 Feb 2009)*
2155 - Xen PVM and KVM have switched the default value for the instance root
2156 disk to the first partition on the first drive, instead of the whole
2157 drive; this means that the OS installation scripts must be changed
2159 - Man pages have been updated
2160 - RAPI has been switched by default to HTTPS, and the exported functions
2161 should all work correctly
2162 - RAPI v1 has been removed
2163 - Many improvements to the KVM hypervisor
2164 - Block device errors are now better reported
2165 - Many other bugfixes and small improvements
2171 *(Released Mon, 26 Jan 2009)*
2173 - Version 2 is a general rewrite of the code and therefore the
2174 differences are too many to list, see the design document for 2.0 in
2175 the ``doc/`` subdirectory for more details
2176 - In this beta version there is not yet a migration path from 1.2 (there
2177 will be one in the final 2.0 release)
2178 - A few significant changes are:
2180 - all commands are executed by a daemon (``ganeti-masterd``) and the
2181 various ``gnt-*`` commands are just front-ends to it
2182 - all the commands are entered into, and executed from a job queue,
2183 see the ``gnt-job(8)`` manpage
2184 - the RAPI daemon supports read-write operations, secured by basic
2185 HTTP authentication on top of HTTPS
2186 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2187 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2188 to migrate to DRBD 8 first while still running Ganeti 1.2)
2189 - DRBD devices are using statically allocated minor numbers, which
2190 will be assigned to existing instances during the migration process
2191 - there is support for both Xen PVM and Xen HVM instances running on
2193 - KVM virtualization is supported too
2194 - file-based storage has been implemented, which means that it is
2195 possible to run the cluster without LVM and DRBD storage, for
2196 example using a shared filesystem exported from shared storage (and
2197 still have live migration)
2203 *(Released Tue, 13 Jan 2009)*
2205 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2206 - Reuse the old instance mac address by default on instance import, if
2207 the instance name is the same.
2208 - Handle situations in which the node info rpc returns incomplete
2210 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2211 - Improved version of batcher:
2213 - state file support
2214 - instance mac address support
2215 - support for HVM clusters/instances
2217 - Add an option to show the number of cpu sockets and nodes in
2219 - Support OSes that handle more than one version of the OS api (but do
2220 not change the current API in any other way)
2221 - Fix ``gnt-node migrate``
2222 - ``gnt-debug`` man page
2223 - Fixes various more typos and small issues
2224 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2230 *(Released Wed, 24 Sep 2008)*
2232 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2233 type of disk exported to fully virtualized instances.
2234 - provide access to the serial console of HVM instances
2235 - instance auto_balance flag, set by default. If turned off it will
2236 avoid warnings on cluster verify if there is not enough memory to fail
2237 over an instance. in the future it will prevent automatically failing
2238 it over when we will support that.
2239 - batcher tool for instance creation, see ``tools/README.batcher``
2240 - ``gnt-instance reinstall --select-os`` to interactively select a new
2241 operating system when reinstalling an instance.
2242 - when changing the memory amount on instance modify a check has been
2243 added that the instance will be able to start. also warnings are
2244 emitted if the instance will not be able to fail over, if auto_balance
2246 - documentation fixes
2247 - sync fields between ``gnt-instance list/modify/add/import``
2248 - fix a race condition in drbd when the sync speed was set after giving
2249 the device a remote peer.
2255 *(Released Tue, 22 Jul 2008)*
2257 - note: the allowed size and number of tags per object were reduced
2258 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2259 - fixed twisted 8.x compatibility
2260 - fixed ``gnt-instance replace-disks`` with iallocator
2261 - add TCP keepalives on twisted connections to detect restarted nodes
2262 - disk increase support, see ``gnt-instance grow-disk``
2263 - implement bulk node/instance query for RAPI
2264 - add tags in node/instance listing (optional)
2265 - experimental migration (and live migration) support, read the man page
2266 for ``gnt-instance migrate``
2267 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2268 has some small improvements in handling its state file
2274 *(Released Fri, 13 Jun 2008)*
2276 - Experimental readonly, REST-based remote API implementation;
2277 automatically started on master node, TCP port 5080, if enabled by
2278 ``--enable-rapi`` parameter to configure script.
2279 - Instance allocator support. Add and import instance accept a
2280 ``--iallocator`` parameter, and call that instance allocator to decide
2281 which node to use for the instance. The iallocator document describes
2282 what's expected from an allocator script.
2283 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2284 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2285 node is lost there is still enough memory to fail over the instances
2287 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2288 ``gnt-cluster verify``, to check for site-specific compliance. All the
2289 hooks will run, and their output, if any, will be displayed. Any
2290 failing hook will make the verification return an error value.
2291 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2292 primary and secondary interfaces
2293 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2294 that is still declared as part of the cluster and has failed.
2295 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2296 specifying output fields, that just adds the chosen fields to the
2298 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2299 export from the filesystem.
2300 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2301 have been added. Using them you can enable/disable acpi and pae
2302 support, and specify a path for a cd image to be exported to the
2303 instance. These parameters as the name suggest only work on HVM
2305 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2306 PAE support will be set to the previously hardcoded values, but the
2307 (previously hardcoded) path to the CDROM ISO image will be unset and
2308 if required, needs to be set manually with ``gnt-instance modify``
2310 - The address to which an instance's VNC console is bound is now
2311 selectable per-instance, rather than being cluster wide. Of course
2312 this only applies to instances controlled via VNC, so currently just
2313 applies to HVM clusters.
2319 *(Released Mon, 18 Feb 2008)*
2321 - more tweaks to the disk activation code (especially helpful for DRBD)
2322 - change the default ``gnt-instance list`` output format, now there is
2323 one combined status field (see the manpage for the exact values this
2325 - some more fixes for the mac export to hooks change
2326 - make Ganeti not break with DRBD 8.2.x (which changed the version
2327 format in ``/proc/drbd``) (issue 24)
2328 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2329 template, allowing migration from DRBD0.7+MD to DRBD8
2335 *(Released Wed, 30 Jan 2008)*
2337 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2339 - add command aliases infrastructure and a few aliases
2340 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2341 man pages and the ``--help`` option of ``gnt-node
2342 list``/``gnt-instance list``
2343 - fix ``gnt-backup list`` with down nodes (issue 21)
2344 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2345 - fix the dist archive and add a check for including svn/git files in
2347 - some developer-related changes: improve the burnin and the QA suite,
2348 add an upload script for testing during development
2354 *(Released Wed, 16 Jan 2008)*
2356 - experimental HVM support, read the install document, section
2357 "Initializing the cluster"
2358 - allow for the PVM hypervisor per-instance kernel and initrd paths
2359 - add a new command ``gnt-cluster verify-disks`` which uses a new
2360 algorithm to improve the reconnection of the DRBD pairs if the device
2361 on the secondary node has gone away
2362 - make logical volume code auto-activate LVs at disk activation time
2363 - slightly improve the speed of activating disks
2364 - allow specification of the MAC address at instance creation time, and
2365 changing it later via ``gnt-instance modify``
2366 - fix handling of external commands that generate lots of output on
2368 - update documentation with regard to minimum version of DRBD8 supported
2374 *(Released Tue, 4 Dec 2007)*
2376 - Log the ``xm create`` output to the node daemon log on failure (to
2377 help diagnosing the error)
2378 - In debug mode, log all external commands output if failed to the logs
2379 - Change parsing of lvm commands to ignore stderr
2385 *(Released Wed, 28 Nov 2007)*
2387 - Another round of updates to the DRBD 8 code to deal with more failures
2388 in the replace secondary node operation
2389 - Some more logging of failures in disk operations (lvm, drbd)
2390 - A few documentation updates
2397 *(Released Tue, 13 Nov 2007)*
2399 - Change configuration file format from Python's Pickle to JSON.
2400 Upgrading is possible using the cfgupgrade utility.
2401 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2402 faster replace disks and is more stable (DRBD 8 has many improvements
2403 compared to DRBD 0.7)
2404 - Added command line tags support (see man pages for ``gnt-instance``,
2405 ``gnt-node``, ``gnt-cluster``)
2406 - Added instance rename support
2407 - Added multi-instance startup/shutdown
2408 - Added cluster rename support
2409 - Added ``gnt-node evacuate`` to simplify some node operations
2410 - Added instance reboot operation that can speedup reboot as compared to
2412 - Soften the requirement that hostnames are in FQDN format
2413 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2415 - Removed dependency on debian's patched fping that uses the
2416 non-standard ``-S`` option
2417 - Now the OS definitions are searched for in multiple, configurable
2418 paths (easier for distros to package)
2419 - Some changes to the hooks infrastructure (especially the new
2420 post-configuration update hook)
2421 - Other small bugfixes
2423 .. vim: set textwidth=72 syntax=rst :